# keyword.py - keyword expansion for mercurial

# $Id$

'''keyword expansion hack against the grain of a DSCM

This extension lets you expand RCS/CVS-like keywords in a Mercurial

repository.

There are many good reasons why this is not needed in a distributed

SCM, still it may be useful in very small projects based on single

files (like LaTeX packages), that are mostly addressed to an audience

not running a version control system.

Supported $keywords$ and their $keyword: substition $ are:

    Revision: changeset id

    Author:   short username

    Date:     %Y/%m/%d %H:%M:%S [UTC]

    RCSFile:  basename,v

    Source:   /path/to/basename,v

    Id:       basename,v csetid %Y/%m/%d %H:%M:%S shortname

    Header:   /path/to/basename,v csetid %Y/%m/%d %H:%M:%S shortname

Simple setup in hgrc:

    # enable extension

    hgext.keyword =

    # or, if script not in hgext folder:

    # hgext.keyword = /full/path/to/script

    # filename patterns for expansion are configured in this section

    [keyword]

    **.py = expand

    ...

'''

from mercurial import context, util

import os.path, re, sys, time

re_kw = re.compile(

        r'\$(Id|Header|Author|Date|Revision|RCSFile|Source)[^$]*?\$')

def kwexpand(mobj, kwfctx):

    '''Called by kwfilelog.read and pretxnkw.

    Expands keywords according to file context.'''

    return '$%s: %s $' % (mobj.group(1), kwfctx.expand(mobj.group(1)))

def kwfmatches(ui, repo, files):

    '''Selects candidates for keyword substitution

    configured in keyword section in hgrc.'''

    files = [f for f in files if not f.startswith('.hg')]

    if not files:

        return []

    candidates = []

    kwfmatchers = [util.matcher(repo.root, '', [pat], [], [])[1]

            for pat, opt in ui.configitems('keyword') if opt == 'expand']

    for f in files:

        for mf in kwfmatchers:

            if mf(f):

                candidates.append(f)

                break

    return candidates

def utcdate(date):

    '''Returns hgdate in cvs-like UTC format.'''

    return time.strftime('%Y/%m/%d %H:%M:%S', time.gmtime(date[0]))

class kwfilectx(context.filectx):

    '''

    Provides keyword expansion functions based on file context.

    '''

    def __init__(self, repo, path, changeid=None, fileid=None, filelog=None):

        context.filectx.__init__(self, repo, path, changeid, fileid, filelog)

    def Revision(self):

        return str(self.changectx())

    def Author(self):

        return util.shortuser(self.user())

    def Date(self):

        return utcdate(self.date())

    def RCSFile(self):

        return os.path.basename(self._path)+',v'

    def Source(self):

        return self._repo.wjoin(self._path)+',v'

    def Header(self):

        return ' '.join(

                [self.Source(), self.Revision(), self.Date(), self.Author()])

    def Id(self):

        return ' '.join(

                [self.RCSFile(), self.Revision(), self.Date(), self.Author()])

    def expand(self, kw):

        '''Called from kwexpand, evaluates keyword.'''

        return eval('self.%s()' % kw)

def reposetup(ui, repo):

    from mercurial import filelog, revlog

    if not repo.local():

        return

    class kwrepo(repo.__class__):

        def file(self, f):

            if f[0] == '/':

                f = f[1:]

            return filelog.filelog(self.sopener, f, self, self.revlogversion)

    class kwfilelog(filelog.filelog):

        '''

        Superclass over filelog to customize it's read, add, cmp methods.

        Keywords are "stored" unexpanded, and expanded on reading.

        '''

        def __init__(self, opener, path, repo,

                     defversion=revlog.REVLOG_DEFAULT_VERSION):

            super(kwfilelog, self).__init__(opener, path, defversion)

            self._repo = repo

            self._path = path

        def iskwcandidate(self, data):

            '''Decides whether to act on keywords.'''

            return (kwfmatches(ui, self._repo, [self._path])

                    and not util.binary(data))

        def read(self, node):

            '''Substitutes keywords when reading filelog.'''

            data = super(kwfilelog, self).read(node)

            if self.iskwcandidate(data):

                kwfctx = kwfilectx(self._repo, self._path,

                            fileid=node, filelog=self)

                return re_kw.sub(lambda m: kwexpand(m, kwfctx), data)

            return data

        def add(self, text, meta, tr, link, p1=None, p2=None):

            '''Removes keyword substitutions when adding to filelog.'''

            if self.iskwcandidate(text):

                text = re_kw.sub(r'$\1$', text)

            return super(kwfilelog, self).add(text,

        def cmp(self, node, text):

            '''Removes keyword substitutions for comparison.'''

            if self.iskwcandidate(text):

                text = re_kw.sub(r'$\1$', text)

            return super(kwfilelog, self).cmp(node, text)

    filelog.filelog = kwfilelog

    repo.__class__ = kwrepo

    # make pretxncommit hook import kwmodule regardless of where it's located

    for k, v in sys.modules.iteritems():

        if v is None:

            continue

        if not hasattr(v, '__file__'):

            continue

        if v.__file__.startswith(__file__):

            mod = k

            break

    else:

        sys.path.insert(0, os.path.abspath(os.path.dirname(__file__)))

        mod = os.path.splitext(os.path.basename(__file__))[0]

    ui.setconfig('hooks', 'pretxncommit.keyword', 'python:%s.pretxnkw' % mod)

    del mod

def pretxnkw(ui, repo, hooktype, **args):

    '''pretxncommit hook that collects candidates for keyword expansion

    on commit and expands keywords in working dir.'''

    from mercurial.i18n import gettext as _

    # above line for backwards compatibility; can be changed to

    #   from mercurial.i18n import _

    # some day

    from mercurial import cmdutil, commands

    if hooktype != 'pretxncommit':

        return True

    cmd, sysargs, globalopts, cmdopts = commands.parse(ui, sys.argv[1:])[1:]

    if repr(cmd).split()[1] in ('tag', 'import_'):

        return False

    files, match, anypats = cmdutil.matchpats(repo, sysargs, cmdopts)

    modified, added = repo.status(files=files, match=match)[:2]

    for f in kwfmatches(ui, repo, modified+added):

        data = repo.wfile(f).read()

        if not util.binary(data):

            kwfctx = kwfilectx(repo, f, changeid=args['node'])

            data, kwct = re_kw.subn(lambda m: kwexpand(m, kwfctx), data)

            if kwct:

                ui.debug(_('overwriting %s expanding keywords\n' % f))

                repo.wfile(f, 'w').write(data)